Department of Linguistics, University of Hawaii, Honolulu ...Although the definitions of Hall (57)—that a pidgin is a language with "sharply reduced" grammatical structure and vocabulary,

Ann. Rev Anthropol 1976 5169-93Copyright O 1976 by Annual Reviews Inc All nghts reserved

PIDGIN AND CREOLE STUDIES • 9575

Derek BickertonDepartment of Linguistics, University of Hawaii, Honolulu, Hawaii 96822

INTRODUCTION

Since pidgin and Creole languages have not previously been surveyed in thisreview series (nor, save as a sub-subdepartment of linguistics, in its biennialpredecessor), it may be appropriate to begin by very briefly summarizing thehistory and development of the field before proceeding to discuss the work thatis currently taking place therein.

Until relatively recently, pidgin and Creole languages were regarded, even bymost linguists, as constituting objects hardly worthy of attention from seriousstudents of language. Despite the fact that attempts to describe such languagesdate back at least to the second half of the eighteenth century (90), and that a fewnineteenth-century linguists, in particular Schuchardt (115), had observed theirpossible relevance to any general theory of linguistic change, the popular viewthat they constituted merely "corrupted" versions of European languages waswidely accepted. As a result, development of the field was delayed, and when itcame was very uneven. According to a survey by Hancock in the Hymescollection (72), there exist at present over 200 pidgin and Creole languages (59),but of these, only about six could be said to have acquired an extensive literature(Haitian Creole, Sranan, Papiamentu, Jamaican Creole, Hawaiian Kdgin-Cre-ole, and Neo-Melanesian or Tokpisin), while many are known only throughanecdotal reference and have never been described at all. Similarly, there hasnever been complete agreement even on the precise boundaries of the field.Although the definitions of Hall (57)—that a pidgin is a language with "sharplyreduced" grammatical structure and vocabulary, native to none of its users,while a Creole is a pidgin that has acquired native speakers—would probably stillbe accepted by a majority of linguists, we will find that more recently somelinguists have tried to narrow the first definition and others to broaden thesecond, while still others, adopting what has been called the "domestic" theoryof Creole origins, have attempted to short-circuit Hall's cycle.

As mentioned in the previous paragraph, early interest in pidgins and Creolescentered around their origins and the extent to which—at a time when theStammbaum theory of genetic relationships was ascendant—they might provecounterexamples to such a theory. Adherents of the theory claimed, for in-

169

170 BICKERTON

Stance, that there could be no such thing as a "mixed language," and yet pidginsand Creoles seemed to show signs of precisely such a mixture; although, in thecase of at least the best-known examples, vocabulary was preponderantly(-F90%) drawn from the Indo-European parent, the syntax seemed to contain anumber of non-IE features. Was it the case that [as Sylvain (127) argued forHaitian Creole] a Creole language was simply the grafting of a European lexiconon an African grammar'' Or did the European component outweigh all otherson every linguistic level? While a number of wnters (66, 128-130, 140) arguedfor at least a modified version of the former position, the majority (41, 55-57,75, 149) continued to maintain the latter view. This was hardly surprising, sinceboth the tradition of Indo-European philology and the currently (i.e. prior to1960) dominant school of structural linguistics both regarded phonology andmorphology as central to the study of language and syntax as relativelyperipheral

However, the nature of the genetic debate was radically changed by theintroduction in the 1960s of a terttum quid m the form of the monogenetichypothesis (122, 132, 150). According to this hypothesis, the similarities foundworldwide among pidgins and Creoles were the result of their having had acommon ancestor, perhaps dating back as far as the medieval Lingua Franca(150), but certainly to an Afro-Portuguese pidgin that is assumed to havedeveloped in fifteenth-century Guinea. Monogenesis, which for all its apparentheterodoxy represents a means of saving many of the assumptions of traditionalhistorical linguistics, entails a belief in relexification (122), the replacement of avocabulary originally Portuguese by English, French, Spanish, or Dutch words,without any effect on other areas of the grammar.

For reasons which remain mysterious to this reviewer, this volume's prede-cessor. The Biennial Review of Anthropology (1959-1971) consistently listedworks dealing with pidgins and Creoles in the section on "Socioiinguistics,"despite the fact that very little work m the field could properly be calledsociolinguistic: an article by Alleyne on language and Jamaican politics (4), acurious essay by Fanon that surprisingly endorses educated French attitudes topidgins and Creoles (42), and an entertaining if rather impressionistic dissertationby Reisman on the ethnography of speaking in Antigua (107) are three of the veryfew pre-1970 examples that spring to mind. For the most part, when it did notseek to delve into origins, work on pidgins and Creoles was purely descriptive,and in general, prestructuralist (e.g 41, 89), early structuralist (25. 37, 54), ortagmemic (95) m orientation; only one study, B. Bailey's (6) analysis of thesyntax of Jamaican Creole, was within the framework of generative grammar.However, during the last few years, several aspects of this picture havechanged, and there has been a considerable rebirth of interest in pidgins andCreoles as possible testing grounds for issues in contemporary theory Threedevelopments in general linguistics have helped to foster this interest. First, therevival of historical linguistics, under generative auspices, created a climatefavorable to the study of the processes of linguistic change. Second, the study oflinguistic variation, long considered minor or irrelevant by most linguists, be-came a legitimate and even respectable field. Third, and perhaps most important

PIDGIN AND CREOLE STUDIES 171

for future studies, the view that languages are inimitably different was largelyreplaced by the view that all languages are fundamentally similar, and this in turnstimulated the search for universais of language.

In consequence, the last 5 years have seen a sharp increase both in the numberof wntings on pidgins and Creoles and the range of topics that such writings havecovered. Although many of these topics overlap, the lack of any single cleardirection in the field makes it necessary, for purposes of review, to subdivide thematerial according to the major centers of interest.

DEFINITIONSIt

A recurring problem in the field has been the precise definition of its boundaries.As numerous writers have pointed out, pidgins and Creoles have traditionallybeen defined in extralinguistic terms, so that (in contrast with the areal fields) it isnot possible to establish the allegiance of any given language simply by lookingat its sound system, grammatical structure, and lexicon, and then comparingthese with those of putatively related languages. At the International Confer-ence on Pidginization and CreoHzation (Jamaica, 1968) a rather fruitless morn-ing was spent discussing "simplification" as a possible formal criterion forpidginization, but "simplification" itself proved quite impossible to define. Inpractice, extralinguistic definitions have been altered and refined. For example.Hall's acceptance, as a pidgin, of any improvised contact language betweenpersons not sharing a common tongue has been challenged by Whinnom (151);according to the latter's view, contact languages produced by a meeting of onlytwo language communities (like the hapa-haole of nineteenth century Hawaii, orthe Italo-Spanish cocoliche of Buenos Aires) represent phenomena too transi-tory and unstable to be classed as distinct languages. The term "pidgin" wouldbe reserved by Whinnom for situations where a reduced or simplified form of Liwas used mainly by native speakers of L2, Ls. L4, etc to communicate with oneanother, rather than with speakers of Li, thus freeing their speech from beingcorrected in the direction of Li and enabling it to develop as an independententity. This view, which would sharply reduce the importance in pidginization ofany deliberate simplification by Li speakers (contrary to the views expressed in19, 44, 96, etc), would probably now be shared by a majority of pidginists.

Another distinction which is sometimes attempted is that between "tradejargons" and pidgins proper (9, 134). It is not clear that such a distinction ismeaningful or can be consistently maintained. It appears to be motivated at leastin part by a recognition that some pidgins are relatively rudimentary and highlyunstable, whereas others are more developed and have relatively homogeneousgrammars. But it is not clear that there is any necessary correlation between thepidgin's function and its linguistic status. Hawaiian pidgin, which was certainlynot a "trade jargon," and indeed gave birth to a Creole, never achieved gram-matical complexity or stability. It would rather seem that all contact languages,whatever their primary function, develop through a series of stages; tradejargons may be likelier to be trapped in the more primitive stages than othertypes of pidgin, but there is no other obvious typological difference.

172 BICKERTON

If the trend in pidgin studies has been to limit the area of application of theterm "pidgin," Creole studies have shown precisely the reverse tendency.Studies by Goodman (51) and Southworth (119) served to raise the possibilitythat pidginization or creolization might have intervened at some stage in thehistorical development of a given language, even where no historical evidencefor such processes has survived. While the cases they present—those of Mbuguand Marathi respectively—seem fairly plausible, there is obviously a danger thatany unusual histoncal change may be "explained" in these terms. Indeed, C JBailey (7, p 134) has come perilously close to equating creolization with lin-guistic change in general

I am taking it for granted that mixtures of systems spoken by native speakers—i.e.Creoles—may occur in different proportions and degrees . Let scientists borrowpainvise from German . and let ii-i e become a productive formative in ordinaryspeech for deriving adverbs from nouns, and this is creolization!

Admittedly, he goes on to state that "one would not wish to speak of creolizationwhere only a few lexical items were borrowed," but the supposition that Creolesare simply any "mixed languages" leads logically to such a position, which ineffect makes "creolization" a redundant term,

A Baileyan view of creolization would, of course, remove Creoles from anynecessary connection with antecedent pidgins; and, indeed, from three othersources have come suggestions that Creoles, properly so described, may existwithout any prior process of pidginization. The first was a paper by Gumperz &Wilson (53), which showed how Marathi, Kannada, and Urdu, as spoken in theIndian village of Kupwar, had undergone so much convergence as to virtuallyshare a common surface syntax, even though the standard forms of theselanguages show many syntactic differences. It was claimed that such conver-gence, based on close contact over an extended period, yielded phenomenaclosely similar to those which characterized creolization, and should thereforebe regarded as special cases of the latter. The second was the suggestion,implicit in Valkoff s work on Portuguese Creoles (144), but made more explicitby Tonkin (135) and Hancock (61), that pidgins themselves may have had aCreole origin. According to this theory, the most likely locus for the origin of anycontact language on the West Coast of Africa lay not in the necessanly fieetingcontacts of traders (in vvhich, as some historical evidence attests, interpretersand even phrase-books were often used) but in the families of lanqados, thoseEuropeans (Portuguese in the first instance, later of other nationalities) whosettled in Guinea and married into various tribal societies. Languages thusdeveloped, it is argued, subsequently became contact languages throughoutWest Africa, and were the ancestors of Caribbean and other Creoles, The thirdsource is a note by Voorhoeve (145) which pointed out a consequence of themonogenetic theory that apparently had not been realized before: that "if thetheory of relexification holds true, a historical Portuguese pidgin has beenrelexified in contact with French masters, without passing through an inter-mediate French pidgin stage." Thus acceptance of monogenesis virtuallyabolishes pidginization as a productive process; one is forced to assume a single


invariant pidgin being transmitted from speaker to speaker just like any otherlanguage.

With respect to all of these definitional proposals, the most one can say is thatthey show how much we still have to learn about linguistic change processes,language transmission, and the various kinds of language-contact situations.Unfortunately, the questions they raise, though of considerable importance, arefar from easy to answer. Many of the types of situation which gave rise to Creoleor creole-like phenomena in the past may not be replicated in the twentiethcentury; those that produced the European-based Creoles—episodes of Westernimperialist expansion—are unlikely ever to be repeated, at least in a similarform. Thus these questions are unlikely to be resolved by empirical study, whilethe only other possible source of solutions, historical reconstruction, is madeextremely difficult by the virtual absence of recorded texts.

At the same time, there is a clear danger that a broadened definition of Creolesmay simply serve to distract attention from what have been traditionally knownas Creoles, i,e. the offspring of pidgin languages. As I shall show in a latersection, there may be reason to believe that these represent differences in kind,rather than in degree, from other kinds of language change, whether contact-generated or internal. If creolization is redefined as no more than massivelinguistic change due to interlingual contact, then these differences may beglossed over, and a potential source of valuable insights into the basic structureof language may be lost.

ORIGINS

As mentioned earlier in this review, the debate about origins has occupied muchof the history of pidgin and creole studies. Though the monogenetic case waswidely accepted in the late 1960s (cf DeCamp 29), this was due more to aprolonged stalemate between previous competing views than to any massivedisplay of supporting evidence. Alleyne (5) pointed out in 1971 that no one hadso far attempted to reconstruct the hypothesized Proto-Pidgin, and this lack stillhas not been remedied. A paper by Voorhoeve (146) which seeks to proverelexification in the case of two of the three Surinam Creoles—Sranan andSaramaccan—represents almost the only recent substantive argument in favorof the monogeneticist position, and even this is not a new argument, but rather agathering of fresh evidence in support of an old one. Indeed, the view thatpidgins and Creoles are predominantly simplifications of their respective super-strates has enjoyed a mild revival (21,68,96,97,143). Work along these lines hasproduced some new evidence, mainly historical, which serves directly or indi-rectly to suggest that deliberate simplification by superstrate speakers mayindeed have existed during early pidginization. However, such work continuesto ignore, downplay, or distort both the number of Creole rules which aredemonstrably nonsuperstratal in origin, and the widespread typological sim-ilarities between Creoles of different genetic affiliation which formed the lynch-pin of the monogeneticist case. Thus it is typical of the "simplificationist"school of thought that it concentrates on superficial morphology rather than

174 BICKERTON

underlying semantics Such writers on French Creoles as Valdman (143) andChaudenson (21), for instance, are quite content to repeat the traditional deriv-ations of Creole tense-aspect markers from French periphrastic constructions,without taking into account that the meanings of the Creole forms by no meansalways match those of the periphrastic forms, though the semantic structure ofFrench Creole tense-aspect systems does match those of at least some non-French Creoles, and these systemic similanjies owe nothing to French or anyother European language.

What has vitiated the origins debate has been an insistence on prematureposition-takmg. For every ten studies that have determined, on the basis of oftensuperficial, partial, and unsystematic evidence, that a given Creole un-questionably belonged to one family or another, we have been lucky to get onewhich eschewed partisan stances and got down to the business of comparison inworkmanlike fashion One of the few pre-1970 examples of the latter genre isGoodman's comparative study of French Creoles (50). Subsequently there havebeen studies by Hancock (60) and Alleyne (still unpublished) of the relationshipsbetween the English-based Atlantic pidgins and Creoles, but no comparablework on Portuguese Creoles apart from a somewhat sketchy and anecdotaltreatment by Valkoff (144) and a bnef article by Ferraz on the Bight of BeninCreoles (45): a comparative study of these four Creoles (Sao Thome, .Angolar,Principe, and Annobon) with the Crioulo of Guinea, Cape Verde Creole, Pa-piamento, and Papia Kristang would seem to be essential, not merely in its ownright, but for the light it might shed on the contention that a Portuguese-basedpidgin was the progenitor of all the European-based Creoles.

Another largely unfilled need is that for comparisons between pidgin or creolelanguages and related non-European languages. Most studies in this area con-centrate on some fairly isolated segment of the grammar such as serial verbs(e.g 11,71, 147, 152), while studies which attempt to embrace a wide variety ofgrammatical phenomena, such as Camden's comparison of South Santo withNew Hebrides Bislama (20), are all too rare. However, two novel forms ofcomparative analysis deserve mention. The first, by Huttar (70), involves takinga list of polysemic root morphemes in a given creole (in this case, Djuka) anddetermining the extent to which their range of meanings is shared by bothputatively related and unrelated languages; for instance, whether, as in Djuka,the morpheme that means '' mouth" is also used to refer to arrowheads and otherpointed objects The results (which Huttar admits are tentative and needconfirmation from other sources) seem cleariy to disconfirm the predictions ofmonogenesis—Creoles with no Afncan substratum scored lower on similaritythan many unrelated indigenous languages—and were hardly more favorable tothose of any "universais" theory. Pidgins and Creoles, as a class, scored onlymarginally higher than a wide sample of indigenous and in the main unrelatedlanguages. Huttar found that the major semantic infiuence, at least at this lexicallevel, came from the substratum: creole languages with an African substratumand some (by no means all) West African languages showed the highest level ofsimilarity. The second, by Lee & Vaughn-Cooke (83). compared Nigerian


pidgin to English and a group of substratal languages on the basis of Chomsky-Hallean feature marking. If a pidgin were really "simpler" than its relatedlanguages, their argument ran, its phonology ought to be less marked; in fact,that of Nigerian pidgin turned out to be more marked, at least with respect to itsphonetic inventory, than some of its contributing languages. Such innovativeapproaches as these will, one hopes, help to break some of the ideologicaldeadlocks which have hampered pidgin-creole studies in the past.

At the same time, one must not neglect the information that can be derivedfrom more traditional studies. Thoiigh historical data on the development ofpidgins and Creoles is notoriously sparse, there can be no doubt that patientgleaning can still yield many more useful facts. Hancock (63) in particular hasstressed the importance of such evidence, although an attempt to provide agenealogy for the Lingua Franca reaching back to Egyptian times contains toomuch speculation and partisan interpretation to serve as a model of its kind.Rather more solid work has been done by Mtihihausler (94) on the early stages ofPacific pidgins; in particular, he has shown the previously unsuspected exis-tence of a pidgin closely related to Tokpisin on German plantations in nine-teenth-century Samoa, and has argued plausibly, if not quite convincingly, forthe role of these plantations in the formation and stabilization of Pacific pidginsgenerally. Evidence has been discovered by Bill Wilson, a graduate student atthe University of Hawaii (reported in 18), that a pidginized Hawaiian antedated,and may well have been the ancestor, of an English pidgin in Hawaii. If, asseems likely, this language, rather than English pidgin, was the plantationlanguage for the first 70 years of sugar cultivation in Hawaii, a number ofcherished sociolinguistic axioms about plantations may have to be rethought.Chaudenson (21), as a by-product of his monumental work on the Reunionlexicon, has shown that one widespread assumption among creolists—that thesimilarities between Mascarene and Caribbean French Creoles could be ex-plained in terms of a common West African pidgin ancestor—cannot be main-tained: too few Africans came to the Indian Ocean, and far too late, for theirspeech to have had more than marginal influence on the Creoles of Reunion,Mauritius, and the Seychelles. Such enquiries, though they cannot themselvesyield an adequate theory of origins, should furnish at least some of the facts onwhich such a theory may eventually be based.

However, the debate about origins has definitely ceased to hold the center ofthe stage, not because it has come anywhere nearer being settled, but becauseinterest has largely shifted to a different, if related, area— the extent to whichlanguage universais, rather than existing individual languages, contribute to thestructure of pidgins and Creoles.

UNIVERSALS

The possibility that there might be some connection between pidgins and Creoleson the one hand and universais oflanguage on the other was voiced as early as1939 by Hjelmslev (67). However, the intellectual climate of the time did not

176 BICKERTON

encourage work on universais. It was not until the present decade that seriousattention was given to the idea.

What was possibly the main impulse came from linguists whose major inter-ests lay outside pidgins and Creoles. Labov, in a widely circulated but never-published paper (81), began to ask questions which related specifically to thefunctional effectiveness of pidgins. If they were reduced or simplified forms oflanguage, how was communication adequately maintained? If pidgins wereadequate for communication, why did Creoles complicate them? These werequestions that had long gone unasked and had badly needed asking: un-fortunately, the data on which Labov based his tentative answers was of aquality far infenor to that of his other studies, and led him to a number ofincorrect conclusions. More explicitly concerned with universais was a paper byKay & Sankoff (76), which circulated in manuscript for 2 years before itsappearance in DeCamp & Hancock's book (31), and which advanced thehypothesis that in a pidgin situation speakers discard constructions from theirown languages which are syntactically marked and are left with a small set ofunmarked structures which show- little or no difference between their deep andsurface forms; in other words, they are able to employ theirfaculte de languageto select a kind of lowest common denominator of simplest forms. Such a lowestcommon denominator would, it was suggested, approximate to the structure of auniversal base. At about the same time, and apparently independently, similarsuggestions were made by a number of younger scholars in the field—Agheyisi(3), Muhlhausler (93), Givon (48)—while Traugott. who had entered the fieldfrom historical linguistics, motivated by an interest in change processes, beganto develop a theory of "natural syntax" which was strongly infiuenced by thework of Labov and Kay & Sankoff (136).

This group of scholars labored under the disadvantage that several of themhad little or no first-hand experience of pidgins, while those who did (Agheyisi,Sankoff) were most familiar with pidgin languages (Nigerian Pidgin, Tokpisin)which had been in existence for a considerable period of time. Yet obviously, ifpidgin speakers did have the power to reduce their language to some kind ofuniversal base, this power would have to be exercised at the beginning, ratherthan the middle or end, of the pidginization process. In fact, such evidence as isobtainable about more primitive pidgins hardly supports the Kay-Sankoff hy-pothesis. In a long and thoughtful article on Chinook jargon, Silverstein (117)showed that its speakers, far from working from any common base structure,rather denved similar surface structures from the distinctive deep structures oftheir own native languages. Nagara (95), analyzing the pidgin English ofJapanese plantation workers in Hawaii, found that much of their phonology andsyntax could be explained in terms of a direct transference of Japanese languagepatterns. Subsequent investigation by Bickerton (18) has confirmed that Hawai-ian pidgin, virtually the only true plantation pidgin which is recoverable today,showed internal differences so gross that it is possible to determine the ethnicityof the speaker from written texts and on grounds of syntax alone. The theorythat pidgm speakers have access to universais cannot, therefore, derive anysupport from empirical studies.


Traugott's picture of a natural syntactic base, derived partly by analogy fromStampian natural phonology (121), is a persuasive one:

If we can accept an unordered semantic base, one which is essentially cognitive, andwhich reflects a kind of semantic weighting . . . then we can argue that a naturalsyntactic process is one which gives spatio-temporally ordered expression to thisunordered cognitive base, in certain restncted ways. I hypothesize, for example, thatthere are natural tendencies to give analytic expression to such grammatical elementsas negation, tense, aspect, mood, logical connectives and so forth (136, p. 315).

However, she errs in attributing to pidgin speakers the capacity to recover sucha level of structure. (I will point out again, at the risk of boring the reader, thataccording to the arguments of both sides, "pidgin speaker" mM.jr here be read as"speaker of a pidgin in the early stages of its formation"; to claim it can alsohave, in this context, its normal meaning of "anyone who speaks a pidginlanguage," and thus include "a speaker of a pidgin language that has beenestablished for several generations," is simply a fudge. That Traugott shares mydefinition is quite clear from her remarks on pp. 318-19 about the capacity tosimplify language.) Linguists often write about pidgins as if people in the originalcontact situation had sat down and said, "We cannot understand one another,therefore let us see if we can devise a pidgin." In fact, pidgin speakers aregenerally under the impression that they are speaking some existing language,albeit in broken form. Typical is an anecdote by Reinecke (104, p. 102) about aChinese laborer in Hawaii, unable to understand the instructions of his newwhite supervisor, who exclaimed "Wasamalla this Haole? He no can speakhaole!" Indeed, such an attitude may persist long after the pidgin has stabilizedand become a creole; a native speaker of Saramaccan, describing to me some ofthe differences be'tween Saramaccan and its more Europeanized neighborSranan, repeatedly referred to the former as "the African language."

If this is the case, then what the pidgin speaker thinks he is doing is trying tolearn an existing language, and the result of his efforts may most profitably becompared with attempts at second language learning under extremely adverseconditions (16, 17). Now, while the thought-experiment pidgin speaker iliesunerringly to universais, the more earthbound foreign language learner hugs hissyntactic home ground closely, and behaves very much as early-stage pidginspeakers do. Indeed, there is something highly counter-intuitive about what isunwittingly implied by Traugott and other members of the "pidgin-universal"school—that the adult speaker's/ocw/?^ de langage is shackled, qua foreignlanguage learner, but completely unbound, qua pidgin speaker. Whether oneagrees or disagrees with the Halle-Lenneberg thesis (58, 88) that languageacquisition is severely inhibited after puberty, there can be no doubt that thelanguage-learning abilities of children are considerably greater than those ofadults. For this reason, it would seem more natural to find access to universaisamong the children of early pidgin speakers, rather than among those speakersthemselves.

The theory that Creoles, rather than pidgins, come closest to language univer-sais has been most explicitly stated by Bickerton (15). This paper suggested that

178 BICKERTON

the similarity of creole tense-aspect systems the world over could be explainedonly by hypothesizing the existence of an innate tense-aspect system, based onhuman cognitive capacities, which surfaced intact, instead of being partiallysuppressed in Stampian, language-particular ways, whenever the input to thechild's language acquisition device failed to find adequate data. Such a situationwould certainly obtain in an early-pidgin plantation comrnunity, where that datawould consist partly of the itself unstable and communicationally inadequatepidgin, partly of a largely unlearnable mix of the previous generation's nativetongues. This theory is still too new and controversial to be satisfactorilyevaluated So far it has been welcomed by Givon (49), whose earlier work (48)had pointed in a similar direction, and Slobin (118), who found support for theunderlying semantic categories proposed in the paper in his own and others'work on child language acquisition; however, it has been criticized by Neff (98),who questions the interpretation of some of the Hawaiian data, and Traugott(137), who finds the umversals proposed to be "overly explicit." Obviously itspredictions must be tested empirically and over as wide a range of languages aspossible. Its most obvious advantage qua theory of creole development (for itsimplications extend to areas outside this field) is that it accounts for and explainsprecisely those facts which have been put forward to justify the monogenetichypothesis.

The debate on umversals, though the newest in the field, seems likely to be themost crucial and far-reaching in years to come. It has already attracted to pidginsand Creoles the attention of a number of specialists from other fields, andprovides an issue that is of potential interest to everyone seriously interested inthe inner mechanics of human language However, only the next few years willdetermine whether it will uncover data rich enough to make its contentionscredible, or, like the "origins" debate before it. degenerate into a theoreticalstalemate, with partisans selecting, out of a broad array of facts, those and onlythose that buttress their own particular case

VARIABILITY

If the study of umversals includes more theory than fact, the same cannot be saidfor another recent development—the study of variation in pidgins and Creoles,particularly in the decreolization process. As noted by Valdman (141), pre-1970orthodoxy had ignored or at least downgraded the amount of variability to befound in these languages. To a large extent this was a political decision, renderedinevitable by popular accusations that they "had no grammar''; thus Hall (57, p.107) felt constrained to argue that " . . . investigations by unprejudiced investi-gators, using modern techniques of linguistic observation and analysis, havedemonstrated conclusively that all pidgins and Creoles, even the simplest, are asamenable to description and formulation as are any other languages." Sincethose "other languages" were supposed to have regular, invariant grammars,pidgins and Creoles must be equally regular if they were to be deemed equallyworthy of study


However, the facts of variation had been noticed at least as early as Rein-ecke's work in the 1930s (104, 105), and were observed in the Caribbeansomewhat later by DeCamp (28). In a penetrating and ahead-of-its-time articleby Stewart (126) that appeared in 1%9, the connection between synchronicvariation and diachronic change was made explicit for the first time. Almostsimultaneously, DeCamp was working out the formalism for implicational scal-ing, which was to become the major operational tool for the variation studies ofthe 1970i(30). His research in Jamaica had indicated that variation was far fromthe chaos which B. Bailey had implied when she wrote that "a given speaker islikely to shift back and forth from Creole to English . . . within a single ut-terance," and that "the lines of demarcation are very hard to draw" (6, p. 1).DeCamp cut this Gordian knot by refusing to draw lines; to him, the "dialectmixture" of somewhere like Jamaica was a "post-creole continuum" with no"structural break" between the furthest creole extreme [which came to beknown as the "basilect,'' a term first used by Stewart (123)] and the form nearestto that of the standard language [described as the "acrolect" in a paper byTsuzaki (138)]. DeCamp claimed that for any linguistic feature found in thecontinuum, its presence in the output of a given speaker would predict thepresence of one set of features, while its absence would predict the absence ofanother set (although presence would make no predictions about absence, andvice versa). The type of table thus produced is illustrated in Table 1:

Table 1

Lects

Form of implicational

F, F2

scales

Features

F3 F4 Fs Fe F7

1

2 - - - - - - X3 - - - - - X +4 - - - - X + +

5 - - - X + + +6 - - X + + + +l - x + + - \ - + +8 x + -(- + + + +9 + + + + + + +__

Table 1 illustrates an exceptionless implicational scale. Minuses indicate theabsence of a given feature; pluses represent its presence. Crosses indicate astate that was not taken into account in DeCamp's original formulation, butwhich subsequent work showed to be necessary, i.e. that in which a speakersometimes uses a given feature and sometimes, even where the opportunityexists, does not. Scales such as this should best be viewed, not as descriptions ofdata, but as abstract measures against which the degree of deviation in actualoutputs can be measured. They make certain predictions about these outputs:for instance. Table 1 predicts that a speaker who lacks feature 5 will also lack

180 BICKERTON

features 1 through 4, and that a speaker who possesses feature 3 will also possessfeatures 4 through 7. However, we can make no predictions about what willoccur to the right of a minus—a speaker lacking feature 5 may categorically lack,or variably or categorically possess, feature 7, for example—and similarly nopredictions can be made to the left of a plus. However, we can say that if aspeaker whose output would otherwise concur with Lect 2 should possessfeature 3, or one who would otherwise occupy Lect 7 should lack feature 6, heviolates the predictions of the scale, and demands either a revision of y;iat scalesuch as will serve to accomodate him, or at the very least some attempt toexplain why he deviates from the majority pattern.

The ontological status and statistical validity of implicational scales have beenthe subject of debate, but there can be no question that they have served to bringto light many linguistic phenomena which were unobserved or inexplicablebefore. In particular, they have shed extensive light on the hitherto puzzlingprocess of "decreolization"—that by which a creole in contact with its super-strate may progressively lose creole charactenstics and eventually come toappear as no more than a rather deviant dialect of that superstrate. In turn, anunderstanding of decreolization has helped to change radically the prevailingopinion about the origins of Black English (see below). It has also been possibleto replicate studies and achieve closely similar results. For instance. Day (26)found a hierarchy of environments for copula deletion in Hawaii identical to thatwhich Labov (79), using a variable-rule format, had found in Black English.Similarly, Washabaugh (148) found that the environments for the replacement ofthe complementizer,^ or fu by tu in Providencia were identical with thosespecified by Bickerton (13) in Guyana. However, his explanation of the phe-nomenon differed in a way that illustrates the limitations, as well as the capa-bilities, of implicational analysis.

Bickerton had argued thatfilfu replacement was determined by three semanticcategories of preceding verb: (a) modals and inceptives, (b) "psychological'"verbs, and (c) all other classes. Washabaugh, however, argued that these cate-gories were poorly motivated—there was no obvious reason why Jilfu re-placement should be determined by them—and that both his and Bickerton'sresults could be explained more parsimoniously in terms of lexical diffusion. It isprobable that in this instance, Washabaugh's analysis is the correct one How-ever, the general conclusion which he draws from this case—that decreolizationis above all a matter of surface forms, and that it is nowhere conditioned by thesemantic level—is hardly tenable in the light of subsequent work by Bickerton(14, 16). The first of these studies shows that the Guyanese mesolectal pronounsystem comes into being by the establishment of an across-the-board genderdistinction which obliterates a preexisting case distinction in the basilect; re-structuring and regularization in the light of semantactic categories, rather thanany mere filtering down of superstrate models, must be the mechanisms oper-ating here. In the second, extensive evidence is given to show that the under-lying Guyanese tense-aspect system goes through several quite complexmutations before it arrives at an approximation to the English system, each


mutation representing a slightly different semantic analysis of the nature ofstates, actions, and events; incidentally, it illustrates one case—the acquisitionof "strong" past forms—where lexical diffusion might have been expected tooperate but where it fails to do so.

The moral of all this is that implicational scales can never determine the formof the grammar. Though they can be used effectively to display the data thatneeds to be accounted for—as has been done by Day (26), Odo (100), Peet (102),and Perlman (103) for parts of the Hawaiian continuum, for instance—and mayserve to suggest grammatical explanations, these suggestions may be misleading(as the case discussed indicates) and in any case need to be supplemented bytraditional methods of linguistic analysis. Awareness of this need to provideadequately motivated explanations is clearly present in the work of youngerCaribbean scholars such as Cooper (22), Roberts (111), and especially Rickford(109, 110), who has demonstrated the interaction of phonological and sem-antactic considerations in the workings of the decreolization process, in par-ticular tracing the disappearance of the doz habitual aspect marker in mesolectalCreoles, which he argues, very convincingly, helped to produce the "distributivebe'" of Black English.

Although most studies of creole variation have used a continuum-type modelbased on DeCamp's original, mention should be made of an alternative ex-planation, suggested by Tsuzaki for Hawaii (139) and at least partially endorsedby Labov (80), i.e. that of "co-existent systems." In fact, no study has yetsucceeded in drawing any kind of satisfactory boundary line between a creoleand its related standard (an apparent exception will be examined in the nextparagraph), or in dividing a creole into any of the subsystems which from time totime have been hypothesized. We must therefore conclude that "co-existentsystems'' are lacking in reality. All that can positively be argued in their favor isthat speakers of Creoles may regard them as "real," i.e. any creole speaker willtell you unhesitatingly whether a given sentence is creole or not (trouble onlystarts when you get two Creole speakers to do this). But one should beware ofconfusing "psychological reality"—a popular rallying cry—with "what the manin the street thinks is going on." If what people do in their daily life contradictswhat they say they do, one should regard as "psychologically real" that whichunderlies their consistent actions, rather than the way those actions may berationalized. In fact, when one asks Creole speakers about their language, one islikely to tap, not some level of intuitive folk wisdom,'but residues of the untruthsabout language taught them in the schools.

It will have been noted that virtually all the evidence for decreolization hasbeen drawn from English-related Creoles. No work of a similar nature has beendone on Portuguese-based Creoles, although the dialect continuum of the CapeVerde islands cries out for such treatment, and of the only two approaches toFrench Creoles from this viewpoint (84,142) the former concluded that there wasno such thing as a continuum between French and Creole. However, Lefebvre'sconclusions still leave doubts. In the first place, they accord rather too neatlywith longstanding establishment beliefs about French-Creole relationships—a

182 BICKERTON

fact that would not matter so much if, prior to the studies described above,similar beliefs had not been held about English-Creole relationships. Secondly,there is some factual evidence (e,g, Valdman 141) that in Haiti, if not in theLesser Antilles, the French-Creole distinction is by no means so sharp. Thirdly,there is the fact that Lefebvre's analysis is based, not on natural speech, but onretellings of a specified folk-tale, in which speakers were directly requested toprovide two versions, one "Creole" and one "French"! One can hardly con-ceive of a methodological framework more loaded in favor of its conclusion, andIt seems likely that any study which (like all the other studies mentioned in thissection) based itself on spontaneous speech in relatively natural settings wouldyield quite different results. In regions where historically related languages arein contact over extended periods, it is hardly plausible to suggest that no resultsshould follow from this fact

Perhaps the most significant insight so far denved from the study of creolevariation is that its synchronic variability may simply represent what mightunder other circumstances have been diachronic change (16, 126). A similarconclusion by Labov with respect to English sound changes (78 and numeroussubsequent works) has helped modify and expand our understanding of histori-cal change in phonology: it seems reasonable to suppose that present and futurestudies of synchronic variation in creole syntax may help to improve our knowl-edge of a still-less-understood area, that of syntactic change.

THE BLACK ENGLISH CONTROVERSY

One specific result of decreolization studies has been a clarification of the debateover the origins of Black English. The traditional position (77, 91, etc) was thatBlack English derived from general English—perhaps with some aberrationsdue to purely "social" causes—and that it showed no influence whatsoever fromAfrican or other sources. This position began to be attacked in the early 1960s bya school of which the most vehement spokesmen were Dillard (32-34) andStewart (124, 125). Their criticisms were twofold: that the "Anglicist" casecould only be supported by positing a wholly random and unprincipled selectionof features from the whole gamut of English dialects (Dillard's "cafeteriaprinciple"), and that it ignored or misanalyzed a number of features which by nostretch of the imagination could be derived from English. The latter featureswere claimed to be clearly of creole origin. However, in the form in which it wasfirst stated, their case won only a limited degree of acceptance. Black Englishwas, in the main, mutually intelligible with White dialects, whereas the Car-ibbean Creoles were not; indeed, if one made a three-cornered comparisonbetween White English, Black English, and the kind of Jamaican Creole de-scribed by B. Bailey (6), there could be little doubt that synchronic BlackEnglish stood closer, on virtually any measure of evaluation, to the former thanto the latter. But at a time when (as mentioned earlier) Creoles were supposed toprove their linguistic respectability by showing that they possessed rules asregular as those of other languages. Bailey and almost all other descriptivists in


the field felt obliged to treat them as unitary systems; moreover, to avoid the slurthat Creoles were merely corrupted versions of standard languages, those sys-tems had to be shown to differ maximally from their superstrates. It followedinevitably that descriptions of Creoles, even Creoles that were really continu-ums, turned out to be descriptions of absolute basilects. There was thus (as far as"Anglicists" were concerned) nothing to bridge the structural gulf betweenCreoles and English except for literary and historical evidence. This the "Cre-olists" had in abundance, but most of it was ruled inadmissible by the "Angli-cists"; ironically, the more liberal attitudes towards race that were thenbecoming prevalent made it easy to dismiss the kind of dialogues found ineighteenth and nineteenth century memoirs, histories, plays, novels etc asattempts to disparage blacks and render them ridiculous, rather than honestefforts to reproduce their speech. A recent example of the power to distract thatthis red herring still retains is given in Fasold's review of Dillard (43). Fasold,who still apparently labors under the misapprehension that any kind of'' simpleand ungrammatical" language is "creole-like," is so worried about "racism"that he manages to miss the whole thrust of Dillard's argument—that theseallegedly "invented" literary forms happen to coincide with actually existingforms in contemporary Caribbean Creoles, about which eighteenth and nine-teenth-century authors, were they never so prejudiced, can hardly have known.However, as "Creolists" have developed their claims by drawing attention tothe role of decreolization—Stewart's already-cited paper (126) is perhaps themost lucid example of this—and as the works listed in the previous section madeclear the workings of the decreolization process, hostility to the "Creolist" caseweakened and in some quarters disappeared. When a semipopular work byDillard appeared and was widely reviewed in 1972 (35), it met with a measure ofacceptance that would have been unthinkable a decade eariier.

The position of Gullah is also critical in this debate. Since Turner's classicstudy (140), no one had seriously disputed that Gullah was a creole. However,Rickford has pointed out (109) that even recent descriptions of Gullah (e.g. 24)treated it on the pattem of other Creoles—i.e. as if only the basilect existed—andthat it was therefore possible to go on regarding it as something quite distinctfrom Black English. However, as has been shown by Rickford's own work, andas will be shown more comprehensively by Stewart's ongoing, but still un-published, work in the Sea Islands, there exists a complete linguistic continuumlinking Gullah with Black English, which closely resembles similar continuumsin the Caribbean and Hawaii.

However, it is likely that a full and satisfactory understanding of the origin anddevelopment of Black English must await a full and satisfactory explanation ofpidgins and Creoles. In an interesting paper by Berdan, favorably disposed to the"Creolist" case (12), the author begins by taking for granted the majorityopinion in the field, that pidgins represent "simplifications" of the superstratelanguage and Creoles represent "complications" of the pidgin. If (as some of theevidence in this review suggests) this view is too simplistic, and should bedropped in favor of the theory that pidgins constitute a grossly handicapped case

184 BICKERTON

of second language learning while Creoles show a partial recourse to the facultede langage (which does not, of course, exclude retention of rules and featuresfrom any of the contributing languages) which later may be subject both tonatural changes and changes due to superstrate pressure, then Berdan's argu-ment, that some Black English constructions such as "relative clause reduc-tion" can be "viewed as simplifications of Standard English grammar," willprove not so much incorrect as irrelevant.

SOCIOLINGUISTICS

Although pidgins and Creoles have traditionally been regarded as somehow"sociolinguistic" in themselves, there was not, prior to the present decade, verymuch work on them that could properly be so termed. In recent years, this lackhas to some extent been remedied. Sociolinguistic studies of Caribbean Creoles—pidgins and Creoles of other areas have hardly even been considered in thislight—fall roughly into two general classes.

The first class is modeled on the Labovian, survey-oriented approach. Workalong these lines has been carried out by Winford in Trinidad (153,154); a muchmore ambitious study has been carried out by Le Page in Belize (85-87). WhileWinford has concentrated mainly on establishing the speech varieties thatcorrelate with class and ethnic divisions, Le Page and his associates have beenmore concerned with the language of the individual as an expression of hisgeneral social allegiance. In the immense linguistic complexity of Belize (wherethe presence of Spanish, Indian ianguages, and Black Carib alongside the usualCaribbean-English continuum provides an unusually wide set of options for thespeaker), each speaker, according to Le Page, "creates the systems for hisverbal behavior so that they shall resemble those of the group or groups, withwhich from time to time he may wish to be identified" (87, p. 2). While thisstatement in essence is correct, the word creates seems hardly apt here; itsuggests that the speaker can actually "make things up," or at least select quitefreely from the whole range of features that the continuum contains. In fact, thestudies described in Section 4 indicate that this is far from being the case. TheCreole menu is not a la carte, it rather consists of a series of "chefs specials,"congenes of features in which if you select one, you must select all. Moreover, itis surely premature to speak of selection until you know what is on the menu; therather inadequate descriptions of past tense and pluralization in Belize creolecontained in (85) leave it quite unclear what is being meant, sociolinguisticatly orin any other way, when a Belizean selects Aw/A:o/n over kom, or ditaiga-dem overtaiga. A similar connection between group allegiance and linguistic level withinthe continuum is shown in Edwards' work on Providencia/San Andres creole(38-40). Here, while again the grammatical analysis may not be above reproach,Edwards was one of the first writers on Creoles to appreciate the implicationalnature of their structure: the fact that any choice of features is constrained by,and in its turn serves to constrain, other choices.

PIDGIN AND CREOLE STXJDIES 185

The second class of sociolinguistic studies concems itself with what Hymeshas called "the ethnography of speaking": roughly, who says what to whom,under what circumstances, and for what purpose. Studies along these lines havebeen produced by Abrahams (1), Abrahams & Bauman (2) and Reisman (108),Such studies are valuable in that they concentrate on something which in strictlylinguistic studies is seldom mentioned, and indeed usually ignored altogether—what speaking in particular ways means to the participants themselves. Goodstudies of this kind can add flesh to the bare bones of formal descriptions, butthey run certain risks, from which the examples mentioned are not exempt. Twosources of danger are the amorphous nature of the material, with its concomitantproblem of what, out of an infinite array of facts, may be relevant and what maynot, and the continuing lack—perhaps inevitable in view of the newness of thefield—of any substantial body of theory. These factors make for mainly anec-dotal treatments; moreover, and perhaps in an attempt to compensate for this,there has been an unfortunate tendency in some studies to return to a di-chotomistic approach, with creole, African-derived and "anti-establishment"expressions, attitudes, etc on one side of the fence, and English, white-oriented,"pro-establishment" expressions, attitudes etc on the other. Granted these twostrands are seen as being interwoven, often in subtle ways, but one feels there ismore to it than that: that the mesolect is no mere artefact of linguistic de-scription , but represents a middle ground in its own right, which (for better or forworse) serves as the home base of many people, perhaps even a majority, in theCaribbean, There are any number of ways, linguistic and nonlinguistic, in whichsuch people can define themselves as distinct and separate both from the Whiteculture which they increasitigly see as exploitative, and the deep creole culturewhich they still, unfortunately, perceive as "low" and "vulgar." Any simpledichotomy means that the life-styles and speech pattems of this important groupare placed outside the scope of analysis.

DESCRIPTIVE STUDIES

The present survey has concentrated on those areas of the field which seemed ofgeneral theoretical interest, and has in consequence mentioned only such de-scriptive works as might have some bearing on current theoretical issues withinthe field. There is no intent to disparage other studies; on the contrary, the fieldsuffers from a lack of straightforward, observationally adequate grammars suchas are taken for granted in many other fields of linguistics. To some extent, thislack is due to the nature of the field; there are (or were until recently) fewerpidginists and creolists than there were pidgins and Creoles to be described, and,to make matters worse, many languages which have been included under thepidgin-creole rubric have only the sketchiest of typological resemblances. Ifone's field is Romance or Polynesian languages, one may reasonably be ex-pected to know something substantive about all the languages within it; to do thesame in pidgin/creole studies demands the competence of a polyglot and the

186 BICKERTON

memory of an elephant! I will therefore merely list, far from exhaustively, whatseem to me to be some of the more interesting descriptive studies that haveappeared over the last few years, several of which are included in the pro-ceedings of the 1975 Honolulu conference (27).

It was tnentioned earlier that one of the problems in the field was the un-availability, synchronically, of certain stages in the pidgin-creole cycle. Whileits social circutnstances do not exactly reproduce those of the classic pidginsituations, the migrations of foreign workers to Germany and Australia haveproduced linguistic phenomena which resemble pidginization. A number ofpapers, mainly unpublished, have attempted to describe the background sit-uation and the types of speech produced (23, 36, 47, 92). The main problem indealing with such phenomena is that we still do not know for certain whether thehne between a "true pidgin" and "foreigner's English" is an unbroken one, orwhether there is a sharp distinction in kind, rather than one of degree (and if so,which side of the line gastarbeiter speech falls); hopefully, more and fullerdescriptions of both gastarbeitersprachen and true pidgins will make detailedcotnparisons possible.

Growing interest in pidgins and Creoles leads to a seemingly never-endingsuccession of new examples coming to light. Among more recent discoveries areptdgins. possibly creolizing or already creolized, in Nagaland, the Sudan, andthe Northerti Territory of Australia. The Australian aboriginal example—RoperRiver Creole, described by Sharpe (116)—is clearly a close relative of Tokpisin,although Its phonology shows a different substratal influence (for instance, it hasno voiceless consonants, so that talk, which serves as the phonetic model for theverb meaning "speak, say" in virtually all Anglo pidgins and Creoles becomesdog rather than the usual tok or taak). The description of Naga pidgin—whichseems to be a blend of a Naga koine with Assamese—by Sreedhar (120) isunfortunately too brief and compressed to give much idea of the language toanyone unfamiliar with its related languages. The same is true of the descriptionsof Juba Arabic, an Arabic-related pidgin spoken in the Southern Sudan, and Ki-Nubi, reportedly a Creole offspring of Juba Arabic, by Bell (10) and Nhial (99).According to Bell (personal communication), however, there are possible sim-ilarities between the tense-aspect system of Ki-Nubi and that ot the European-language-related Creoles. Obviously, pidgins and Creoles which claim no Eu-ropean ancestors present much more of a problem for the Western linguist,especially when, as in both the Naga and Ki-Nubi cases, contributions to thecontact language must have come from at least two distinct language families.However, their importance would be hard to overstress. Any theory aboutpidginization and creolization that is based exclusively on European-influencedmodels cannot but be suspect, especially if it claims universal significance;unfortunately, clear and unambiguous cases of non-European pidgins creolizinghave not so far been proven to exist. It is therefore vital iti the future to collect asmuch data as possible on any situation where it is believed that, prior toEuropean contact, pidginization or creolization may have taken place; so farotily Sango (113) and Chinook Jargon (74, 117) have received more than cursory


attention, apart from work by Heine (64, 65) on pidginized versions of Africanlanguages.

With regard to better-known pidgins and Creoles, some of the more interestingpapers have been concerned with the lexicon. In addition to the paper by Huttaralready cited, one may mention a paper by Frake (46)—unfortunately notfollowed up by subsequent publications—on the remarkable principles accord-ing to which the lexicon of Zamboangueno is divided between words ofPhilippine and Spanish origin; a comparative word-list for Djuka, Sranan, andSaramaccan by Huttar (69) which indicates some fascinatingly divergent routesfollowed by phonological change in the three languages; and a paper by Hancock(62) on "incoining"—his own coinage for the processes by which a Creole canexpand its vocabulary without increasing its stock of loan words.

In phonology, Tinelli (133) has analyzed nasalization in Haitian Creole, P*apen(101) has examined the rules affecting verb suffixes in Mascarene Creoles, andJohnson (73) has described morpheme-structure rules in the Atlantic Anglo-creoles. All these papers show a familiarity with modem methods of gram-matical analysis which is less often shared by the writers on syntax. In that field,one of the most interesting contributions is SankofTs study (112) of the markingof relative clauses in Tokpisin, which illustrates the development of quasi-obligatory grammatical markers out of what were originally purely functionaldiscourse devices (and which even now cannot be satisfactorily describedwithout taking discourse factors such as speaker presupposition into account).Also discourse-oriented was a study of narrative pattems in Saramaccan byGrimes & Glock (52). Since functional discourse characteristics would seem apriori likely to have universal status, studies such as these should help to open upan important new approach to future comparative studies in the field.

One thing which that field conspicuously lacks for any of its languages is thetype of compendious reference grammar represented by e.g. Schachter &Otanes' work on Tagalog (114). A work of this type on any pidgin or Creolewould serve as an invaluable base for the exhaustive cataloguing of similaritiesand dissimilarities which needs to be carried out during the next decade. Mostattempts to produce grammatical overviews of individual pidgins and Creolesdeserve, and many modestly claim, only the title of "sketch," but Tokpisin hasbeen better served than many by two accounts from Laycock (82) and Wurm(155). These grammars, though pedagogically oriented, give good general out-lines of the language and have the added advantage that they contain a number ofinteresting texts; similar remarks would also apply to Baker's (8) study ofMauritian creole.

The field still lacks a good general introduction. Hall's introductory volume(57), which was not above criticism when it appeared nearly 10 years ago [seeTaylor's (131) penetrating review], has now been outdated in very many re-spects by the research of the last decade. However, the only general work tohave appeared since (Todd 134), while it updates many of Hall's conclusions, isunduly brief and selective in its coverage. Fortunately, by the time this reviewappears, the compendious bibliography of the field by Reinecke et cd (106)

188 BICKERTON

should at last be available. This work, with its several thousand annotatedreferences, covers all generally recognized pidgins and Creoles, is virtuallyexhaustive up to 1970, and contains many of the more important titles publishedsubsequently; it should prove an indispensable aid to every scholar in the field.Also of service to present and future scholars should be the Journal of CreoleStudies, under the editorship of Edgar Polome and Ian Hancock, which is alsoscheduled to appear in 1976. This joumai, which will contain articles on literaryand social matters relating to creole studies as well as purely linguistic treat-ments, will supplement but not supplant what has hitherto been the only organdevoted solely to pidgins and Creoles—the Carrier Pidgin, a quarterly newsletterformerly edited by Barbara Robson, in the future to be edited by StanleyTsuzaki and John Reinecke.

SUMMARY AND FORECAST

As the foregoing account should have indicated, the study of pidgins and Creoleshas been passing through a period of rapid growth and diversification. Ten yearsago it was regarded as little more than a quaint backwater of linguistics; now ittends to be treated with interest and some respect, mixed perhaps with ameasure of mild scepticism as to whether it can really deliver all that its morevocal adherents have promised. Only time will tell whether it can maintain itsgrowth rate and achieve its potential, but there can be little doubt that thatpotential is a considerable one. The pidgin/creole field is unique among fields oflinguistic study in that it unites the concem for specific languages of the variousareal fields with an obligation to deal in processes which, if not necessarilyuniversal, are at least more than family-specific. The difficulties which thisinvolves have already been mentioned; one should not underestimate the com-pensating advantages.

However, the field still has a long way to go. This review, as well as chroni-cling its achievements, has tried also to indicate some of the limitations which itmust overcome. It should not be forgotten that its resources, in terms of scholarsand sheer finance, are far less than those of many areal fields, and yet need to bespread over a vaster area of ground. The spread so far has been remarkablyuneven. It is hardly surprising perhaps that of the only three studies of pidgins orCreoles that have obtained financial support on a really generous scale, two havebeen in the United States (in what are almost the only two pidgin-creole areasAmerica has, Hawaii and the Sea Islands) while the third has been in a politicallysensitive dependency of Great Britain (British Honduras). There is nowhereany institute or university department devoted to the study of pidgins or Creoles;there is no chair of pidgin and creole languages in any European university.Considering the potential that the field has for adding to our knowledge of thehuman language faculty, and through that faculty to our knowledge of the humanmind itself, one can but hope that these deficiencies will be speedily remedied.But the remedying of them will in turn depend on the capacity of the field, withinthe next decade, to maintain its present rate of development and prove beyonddoubt the indispensability of its contribution to a complete science of man.


Literature Cited

1. Abrahams, R. D. 1972. The trainingof the man of words in talkingsweet. L/>i5 1:15-29

2. Abrahams, R. D., Bauman, R.1971. Sense and nonsense in St.Vincent; speech behavior and deco-ruiVi in a Caribbean community.Am. Anthropol. ly.ldl-ll

3. Agheyisi, R. N. 1971. West AfricanPidgin English: simplification andsimplicity. PhD thesis. StanfordUniv., Calif.

4. Alleyne, M. C. 1963. Commu-nication and politics in Jamaica.Caribb. Stud. 3:22-^1

5. Alleyne, M. C. 1971. Acculturationand the cultural matrix of cre-olization. See Ref 72, 169-86

6. Bailey, B. 1966. Jamaican CreoleSyntax. Cambridge: C. U. P. 164pp.

7. Bailey, C. J. N. 1973. Vanation andLinguistic Theory. Washington DC:Cent. Appl. Ling. 162 pp.

8. Baker, P. 1972. Kreol, a De-scription of Mauritian Creole. Lon-don: Hurst. 221 pp.

9. Baron, N. S. 1977. Trade jargonsand pidgins: a functionalist ap-proach. See Ref. 27

10. Bell, H. 1977. Pidgin Arabic andthe language survey of the Sudan.See Ref 27

11. Bendix, E. H. 1970. Serial verbs inCreole and West African. Presentedat 69th Ann. Meet. Am. Anthropol.Assoc., San Diego

12. Berdan, R. 1977. Sufficiency condi-tions for a prior creolization ofBlack English. See Ref. 27

13. Bickerton, D. 1971. Inherent vari-ability and variable rules. Found.Lang. 7:457-92

14. Bickerton, D. 1973. The nature of aCreole continuum. Language 49:640-69

15. Bickerton, D. 1974. Creolization,linguistic universals, natural sem-antax and the brain. Work. Pap.Ling. 6.3:125-41. Univ. HawaiiDep. Ling.

16. Bickerton, D. 1975. Dynamics ofa Creole System. Cambridge:C. U. P. 224 pp.

17. Bickerton, D. 1976. Pidginizationand creolization: language acqui-sition and language universals. InPidgin and Creole Linguistics, ed.A. Valdman. Bloomington: IndianaUniv. Press

IB Bickerton, D., Odo, C. 1976 Re-

port on NSF Grant No. GS-39748,I^rt 1: general phonology and pid-gin syntax. Univ. Hawaii (mimeo)

19. Bloomfield, L. 1933. Language.New York: Holt. 566 pp.

20. Camden, W. G. 1977. Parallels instmcture of lexicon and syntax be-tween New Hebrides Bislama andthe South Santo language spoken atTangoa. See Ref. 27

21. Chaudenson, R. 1974. Le Lexiquedu Parler Creole de la Reunion.Paris: Libraire Honore Champion.1249 pp.

22. Cooper, V. O. 1977. On the notionof decreolization and St. Kitts Cre-ole personal pronouns. See Ref. 27

23. Clyne, M. G. 1977. German andEnglish working pidgins. See Ref.27

24. Cunningham, I. A. E. 1970. A syn-tactic analysis of Sea Island Creole(Gullah). PhD thesis. Univ. Michi-gan, Ann Arbor

25. D'Ans, A. M. 1968. Le CreoleFranqais d'Haiti. The Hague: Mou-ton. 181 pp.

26. Day, R. R. 1972. Patterns of vari-ation in copula and tense in theHawaiian post-creole continuum.PhD thesis. Univ. Hawaii, Hon-olulu. 169 pp.

27. Day, R. R., Bickerton, D. 1977.Papers from the International Con-ference on Pidgins and Creoles,Honolulu, 1975. Honolulu: Univ.Hawaii Press. In press

28. DeCamp, D. 1961. Social and geo-graphic factors in Jamaican dialects.In Creole Language Studies, ed.R. B. Le Page, 61-84. London:Macmillan

29. DeCamp, D. 1971. The study ofpidgin and creole languages. SeeRef. 72, lJ-39

30. DeCamp, D. 1971. Towards a gen-erative analysis of a post-creolespeech community. See Ref. 72,349-70

31. DeCamp, D., Hancock, I. A., eds.1974. Pidgins and Creoles: CurrentTrends and Prospects. WashingtonDC: Georgetown Univ. Press. 137

32 Siilard, J. L. 1967. Negro chil-dren's dialect in the inner city. Fla.FL Rep. 5.3:7-10

33. Ibid 1968. Non-standard Negro di-alects—convergence or divergence?6.2:9-12

34. Ibid 1970. Principles in the history

190 BICKERTON

of Amencan English—paradox, vir-gmity and cafeteria. 8.1/2:32-34

35 Dillard, J. L. 1972. Black English-its History and Usage in the UnitedStates New York' Random House.361 pp

36. Dittmar, N., Klein, W 1974. Un-tersuchungen zum Pidgin-Deutschspaniscber and ltalienischer arbeiterm der BRD. Deutsches Seminar derUniversitat Heidelberg (mimeo)

37 Dwyer. D . 1969 Introduction toWest African Pidgin English. EastLansing: Afr. Stud Cent., MichiganState Univ. 572 pp.

38 Edwards. J D. 1968 Social lin-guistics on San Andres and Provi-dencia Islands. Louisiana StateUniv (mimeo)

39 Edwards, J D 1970. Aspects of bi-lingual behavior on San Andres Is-land, Colombia Louisiana StateUniv (mimeo)

40 Edwards. J D 1974. Africaninfluences on the English of SanAndres Island, Colombia See Ref.31, 1-26

41 Fame, J. 1937 Philologie Creole.Port-au-Prince: Impnmene del'Etat. 303 pp.

42. Fanon, F 1952. Le noir et le lan-gage. In Peau Noire, MasquesBlancs. F. Fanon, 33-52. Paris-Editions du Seine. 237 pp.

43 Fasold, R. 1975 Review of Ref 35,Language in Society 4:198-221

44. Ferguson, C. A. 1971 Absence ofcopula and the notion of simplicity:a study of normal speech, baby talk,foreigner talk and pidgin. See Ref.72, 141-50

45. Ferraz. L. 1977. The ongin and de-velopment of four Creoles in theGulf of Guinea. See Ref. 27

46 Frake, C. O. 1971. Lexical originsand semantic structure in PhilippineCreole Spanish. See Ref. 72, 223-42

47. Gilbert, G. G., Orlovic, M. 19/7.Pidgin German spoken by foreignworkers in West Germany: thedefinite article. See Ref. 27

48. Givon. T. 1973. Prolegomena toany creology. Univ. Califomia, LosAngeles (mimeo)

49 Givon, T 1975. Postscript to Ref48 Univ. California, Los Angeles(mimeo)

50. Goodman, M. F. 1964 A Compar-ative Study of French Creole Di-alects The Hague:Mouton 143 pp.

51. Goodman. M F. 1971. The strangecase of Mbugu (Tanzania) See Ref.72. 243-54

52. Grimes, J E , Glock, N. 1970. ASaramaccan narrative pattem. Lan-guage 46:408-25

53. Gumperz, J. J., Wilson, R 1971Convergence and creolization: acase from the Indo-Aryan/Dravidianborder. See Ref. 72, 151-68

54 Hall, R. A. Jr. 1948. The Imguisticstructure of Taki-taki Language24:92-116

55. Hall, R. A. Jr. 1952 Pidgin Englishand linguistic change Lingua3:138-46

56. Hall, R A Jr 1958. Creole lan-guages and genetic relationshipsWord 14:367-73

57 Hall. R. A. Jr 1966. Pidgin andCreole Languages. Ithaca: CornellUniv. Press. 188 pp.

58 Halle, M. 1964. Phonology m gener-ative grammar In The Structure ofLanguage, ed. J A. Fodor. J J.Katz, 334-52. Englewood Cliffs,NJ: Prentice-Hall. 612 pp

59. Hancock, I. A. 1969. A map andlist of pidgin and creole languages.See Ref. 72, 509-23

60. Hancock, I. A. 1969. A provisionalcomparison of the English-derivedAtlantic Creoles. Afr Lang Rev.8:7-72

61 Hancock, 1. A. 1972. A domesticoHgin for the English-derived Atlan-tic Creoles. Fla FL Rep. 10.1/2:7-8, 52

62. Hancock, 1. A. 1975. The use ofhistorical evidence in the re-construction of the genesis and de-velopment of pidgins and Creoles.Univ. Texas, Austin (mimeo)

63. Hancock, I. A. 1977. Lexical ex-pansion within a closed system. SeeRef. 27

64. Heine, B. 1970 Status and Use ofAfrican Lingua Francos. NewYork: Humanities Press. 206 pp.

65. Heine, B. 1977. Some gener-alizations on African-based pidginsSee Ref. 27

66. Herskovits, M. J. 1941. The Mythof the Negro Past. New York:Harper. 374 pp.

67. Hjelmslev, L. 1939. Caracteresgrammaticaux des langues Creoles.Presented at Congr. Int. Sci. An-thropol. Ethnol., Copenhagen

68. Hull. A. 1977 On the origin andchronology of the French-basedCreoles See Ref. 27

69. Huttar, G. L. 1972 A comparativeword list for Djuka In Languagesof the Guianas, ed. J E. Grimes,12-21 Summer Inst Ling. 91 pp.


70. Huttar, G. L. 1975. Sources of Cre-ole semantic structures. Language51:684-95

71. Huttar, G. L 1977. Some Kwa-likefeatures of Etjuka syntax. See Ref.27

72. Hymes, D., ed. 1971. The Pid-ginization and Creolization of Lan-guages. Cambridge: C. U. P. 530pp.

73. Johnson, M. C. 1974. Two mor-pheme-structure rules in an Englishproto-creole. See Ref. 31, 118-29

74. Johnson, S. V. 1977. Chinook Jar-gon variations: towards the Com-pleat Chinooker. See Ref. 27

75. Jourdain, E. 1956. Du Franqais auxParlers Creoles. Paris: Klincksieck.334 pp.

76. Kay. P., Sankoff, G. 1974 A lan-guage-universals approach to pid-gins and Creoles. See Ref. 31, 61-72

77. Krapp, G. P. 1925. The EnglishLanguage in America. New York:Century. 2 vols. 594 pp.

78. Labov, W. 1966. The SocialStratification of English in NewYork City. Washington DC: Cent.Appl. Ling. 655 pp

79 Labov, W. 1969. Contraction, de-letion and inherent variability of theEnglish copula. Language 45:715-62

80. Labov, W. 1971. The notion of 'sys-tem' in Creole languages. See Ref.72, 447-72

81 Labov, W. 1971. On the adequacyof natural languages: I. The devel-opment of tense Univ. Pennsylva-nia (mimeo)

82. Laycock, D. 1970. Materials inNew guinea Pidgin (Coastal andLowlands). Canberra: Aust. NatlUniv. 99 pp.

83. Lee, M. H., Vaughn-Cook, A. F.1977. On inventories and syllable-structures. See Ref. 27

84. Lefebvre, C. 1974. Discretenessand the linguistic continuum in Mar-tinique Anthropol. Ling. 16.2:47-78

85. Le Page, R. B. 1972. Preliminaryreport on the sociolinguistic surveyof Cayo District, British Honduras.Lang. Soc. 1:155-72

86. Le Page, R. B. 1973. The conceptof competence in a creole/contactsituation. York Pap. Ling. 3:31-50

87 Le Page, R. B., Christie, P., Jur-dant, B., Weekes, A. J., Tabouret-Keller, A. 1974. Further report onthe sociolinguistic survey of multi-lingual communities: survey ofCayo District, British Honduras.Lang. Soc. 3:1-32

88. Lenneberg, E. H. 1967. BiologicalFoundations of Language. NewYork: Wiley. 489 pp.

89. Lenz, R. 1928. El Papiamento, laLengua Criolla de Curazao. San-tiago de Chile: Balcells. 341 pp.

90. Magens, J. M. 1770. Grq/nmaticaover det Creolske Sprog. Co-penhagen: Gerhard Giese Salikath

91 McDavid, R. I. Jr., McDavid,V. G. 1951 The relationship of thespeech of American Negroes to thespeech of Whites. Am. Speech26:3-17

92. Meisel, J. M. 1975 The language offoreign workers in Germany.Presented at 4th Int. Congr. Appl.Ling., Stuttgart

93. Muhlhausler, P. 1972. Pidginizationand simplification of language.MPhil thesis. Univ. Reading, Eng-land. 179 pp.

94. Muhlhausler, P. 1975. SamoanPlantation Pidgin English and theorigin of New Guinea Pidgin.Presented at 1975 Ann. Conf Ling.Soc. Australia, Canberra

95. Nagara, S. 1972. Japanese PidginEnglish in Hawaii- a Bilingual De-scription. Honolulu: Univ. HawaiiPress. 322 pp.

96. Naro, A. 1973. The origin of WestAfrican Pidgin. Presented at 9thReg. Meet. Chicago Ling. Soc ,Chicago

97. Naro, A. 1977. The origin of PidginPortuguese. See Ref. 27

98. Neff. K. J. 1977. Go m HawaiianCreole. See Ref. 27

99. Nhial, A. A. J. 1977. Ki-Nubi andJuba Arabic: the relationship be-tween a Creole and a pidgin of theArabic language See Ref. 27

100. Odo, C. 1970. English patterns inHawaii. Am. Speech 45:234-39

101. Papen. R. A. 1977. Nana k nana,nana k napa or the strange case of'e-deletion' verbs in Indian OceanCreole. See Ref. 27

102. Peet, W. Jr. 1973. Omission of sub-ject relative pronouns in HawaiianEnglish restrictive relative clauses.In Towards Tomorrow's Linguistics,ed. R W. Shuy, C.-J. N. Bailey.253-66. Washington DC: George-town Univ Press

103 Perlman, A. M. 1973. Grammaticalstructure and style-shift in HawaiianPidgin and Creole. PhD thesis.Univ. Chicago, 111 284 pp.

104. Reinecke, J. 1969. Language andDialect in Hawaii Honolulu: Univ.Hawaii Press. 254 pp.

192 BICKERTON

105 Reinecke, J., Tokimasa, A. 1934.The English dialect of Hawaii. AmSpeech 9:48-58, 122-31

106. Reinecke, J., DeCamp, D., Han-cock, I. A., Tsuzaki, S., Wood, R1975. A Bibliography of Pidgins andCreoles. Honolulu. Univ. HawaiiPress

107. Reisman, K. M. L. 1965. The isle iifull of noises - a study of Creole inthe speech patterns of Antigua,West Indies. PhD thesis. HarvardUniv . Cambridge, Mass.

108. Reisman, K. M L. 1977. Remarkson conflict and meaning in 'continu-ous' systems. See Ref. 27

109. Rickford. J R. 1974 The insightsof the mesolect See Ref. 31, 92-117

110. Rickford, J. R. 1977. How does dozdisappear'> See Ref. 27

111. Roberts. P A. 1977. The impor-tance of certain linguistic theoriesfor showing important relationshipsin a Creole language. See Ref 27

112. Sankoff. G . Brown. P. 1975. Onthe origins of syntax in discourse, acase study of Tok Pisin relativesUniv. Montreal. 86 pp (mimeo)

113. Samarin, W J. 1967 A Grammar ofSango. The Hague: Mouton. 284PP

114. Schachter, P., Otanes. Fe T. 1972Tagalog Reference Grammar. Ber-keley Univ Califomia Press. 566

115. Schuchardt, H E. M. 1909. Die hn-gua franca. Z. Rom. Philol 33"441-61

116. Sharpe. M. C. 1974. Notes on the'Pidgin English" Creole of RoperRiver Aust Inst. Aboriginal Stud.Newsl. (new ser.) 2.2-12

117 Siiverstein, M. 1972. Chinook jar-gon: language contact and the prob-lem of multilevel generative sys-tems Language 48' 378-406,596-625

118 Slobin. D. I. 1975. The more itchanges. . . On understanding lan-guage by watching it move throughtime. Presented at Stanford ChildLanguage Forum, Stanford, Calif.

119 Southworth, F. C. 1971. Detectingpnor creolization: an analysis of thehistorical origins of Marathi. SeeRef. 72, 255-74

120. Sreedhar, M. V. 1977. Standard-ization of Naga Pidgin. See Ref. 27

121. Stampe, D. 1972 A dissertation onnatural phonology. PhD thesis.Univ Chicago. III. 79 pp.

122 Stewart. W A. 1962 Creole lan-

guages in the Caribbean. In Studyof the Role of Second Languages inAsia, Africa and Latin America, ed.F. A. Rice, 34-53. Washmgton DC:Cent. Appl. Ling. 128 pp.

123. Stewart, W A. 1965. Urban negrospeech: sociolinguistic features af-fecting English teaching. In SocialDialects and Language Learning,ed. R. W. Shuy, 10-18. Champaign,111.: Natl Counc. Teach. EngT. 157pp.

124. Stewart, W A. 1967. Sociolinguisticfactors in the history of AmericanNegro dialects. Fla FL Rep5.2:11-30

125. Ibid 1968. Continuity and change inAmerican Negro dialects. 6.1:3-18

126. Stewart, W. A. 1969. Historical andstructural bases for recognition ofNegro dialect. Proc. 20th AnnRound Table, Georgetown Univ ,239-47

127. Sylvain, S C 1936. Le CreoleHaitien- Morphotogie et SyntaxePort-au-Prince: Puillet. 180 pp.

128 Taylor, D. 1959. On function versusform in 'non-traditional' languagesWord 15:485-89

129. Taylor, D. 1960. Language shift orchanging relationship? Int. J. Am.Ling. 26:144-61

130. Taylor. D. 1963. The ongin of WestIndian Creole languages: evidencefrom grammatical categories AmAnthropol. 65:800-14

131. Taylor, D. 1%7. Review of Ref 57.Language 43:817—24

132. Thompson, W. A 1961. A note onsome possible affinities between theCreole dialects of the old world andthose of the new. In Creole Lan-guage Studies, ed R B Le Page.107-13 London: Macmillan 130 pp.

133 Tinelli, H. 1974 Generative andcreolization processes- nasality inHaitian Creole. Lingua 33:344-66

134. Todd, L. 1974. Pidgins and Creoles.London. Routledge & Kegan Paul106 pp

135. Tonkin, E. 1971. Some coastal pid-gins in West Afnca. In Social An-thropology and Linguistics, ed. E.Ardener. 129-55. London: Tavi-stock. 318 pp.

136. Traugott, E. C. 1973. Somethoughts on natural syntactic pro-cesses In New Ways of AnalyzingVariation in English, ed. C -J. NBailey, R. W. Shuy, 31^22. Wash-ington DC; Georgetown Univ.Press. 373 pp.


137. Traugott, E. C. 1976. Pidginization,creolization and language change. InPidgin and Creole Linguistics, ed.A. Valdman. Bloomington: IndianaUniv. Press

138. Tsuzaki, S. M. 1966. HawaiianEnglish: pidgin, creole or dialect?Pac. Speech 1.2:25-28

139. Tsuzaki, S. M. 1971. Co-existentsystems in language variation: thecase of Hawaiian English. See Ref72, 327-39

140. Tumer, L. D. 1949. Africanisms inthe Gullah Dialect. Chicago: Univ.Chicago Press. 317 pp.

141. Valdman, A. 1969. The languagesituation in Haiti. In Research Re-sources of Haiti, ed. R. Schaedel,155-203. New York: Res. Inst.Study of Man. 624 pp.

142. Valdman, A. 1973. Some aspects ofdecreolization in Creole French.Current Trends in Linguistics, 11;507-36. The Hague: Mouton. 604pp.

143. Valdman, A. 1977. A pidgin originfor Creole French? See Ref. 27

144. Valkoff, M. F. 1966." Studies inPortuguese and Creole. Johannes-burg: Witwatersrand Univ. Press.282 pp

145. Voorhoeve, J. 1971. A note on re-duction and expansion in grammarSee Ref. 72, 189 .

146. Voorhoeve, J 1973. Histodcal and

linguistic evidence in favor of therelexification theory in the for-mation of Creoles. Lang. Soc.2:133-46

147. Voorhoeve, J. 1977. Serial verbs inCreole. See Ref. 27

148. Washabaugh, W. 1974. Variabilityin decreolization on Providence Is-land, Colombia. PhD thesis. WayneState Univ., Detroit, Mich.

149. Weinreich, U. 1958. On the com-patibility of genetic relationship andconvergent development. Word14:374-79

150. Whinnom, K. 1965. The origin ofthe European-based pidgins and Cre-oles. Orbis 14:509L-27

151. Whinnom, K. 1971. Linguistic hy-bridization and the 'special case' ofpidgins and Creoles. See Ref. 72,91-115

152. Williams, W. R. 1971. Serial verbconstructions in Krio. Stud. Afr.Ling., Suppl. 2:47-^5

153. Winford, D. 1972. A sociolinguisticdescription of two communities inTrinidad. PhD thesis. York Univ.,England

154. Winford, D. 1977. The creole sit-uation in the context of so-ciolinguistic studies. See Ref. 27

155. Wurm, S. A. 1971 New GuineaHighlands Pidgin: Course Materi-als Canberra: Aust. Natl Univ. 175pp.

Documents

Department of Linguistics, University of Hawaii, Honolulu ...Although the definitions of Hall (57)—that a pidgin is a language with "sharply reduced" grammatical structure and vocabulary,